Three statistical methods of analyzing repeated measurements on microbiological changes in plaque were investigated. The simplest method to use is the univariate analysis of variance. This approach has an exact F distribution when hypotheses are tested under restrictive assumptions of normality and equal variances and covariances. An alternative approach called profile analysis is based on the multivariate analysis of variance technique. In the profile analysis, the readings on each subject at t times are regarded as a t-dimensional vector. A third approach is based on the polynomial growth curve model, i.e., a separate polynomial growth model is fit for each group of data. Data from an eight-week study on the natural transmission of S. mutans in rats on five different dietary concentrations of sucrose was used to illustrate these methods. Statistical analysis indicated that the assumption of equal variance and covariance in the univariate analysis of variance was violated. It was shown that different sucrose concentrations in the diets produced significantly different S. mutans growth profiles. Each sucrose group over time was summarized by a single polynomial equation of time.